You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
So I'm using this over here on the cluster at EMBL (with my own ~/.batchjobs.R file) and it works perfectly except for one thing - I had to modify it to use an alternative tempdir() function, because our cluster needs to write things to a different directory ('scratch'), and will fail if given a '/tmp/' path.
The text was updated successfully, but these errors were encountered:
Happy to see this gets used over at the main site as well! 👍
tempdir() respects the TMPDIR environment variable. If you set it to /scratch/... in your environment, a call to tempdir() will return a directory there instead of /tmp - so there is no need to change the call.
Note that we will phase out BatchJobs in favour of an rzmq solution because the former starts an R session per function call, which kills the file system for a lot of short-running calls (see this bug report and the hpc_2.0 branch). After this change, there will be no more temporary directories.
Should have thought of that, thanks Michael - and yes, konrad gave me access. Can I invite other predocs here? I actually think we should vastly expand what you guys have done here. Even access to other people's buggy, uncommented code would have saved me a lot of time during my PhD - e.g. code for ggplot2 plots.
So I'm using this over here on the cluster at EMBL (with my own ~/.batchjobs.R file) and it works perfectly except for one thing - I had to modify it to use an alternative
tempdir()
function, because our cluster needs to write things to a different directory ('scratch'), and will fail if given a '/tmp/' path.The text was updated successfully, but these errors were encountered: